Last week, we noted an attempt to fix a hardware problem with software, which backfired pretty dramatically for Ford when they tried to counter the tendency for driveshafts to fall out of certain of their cars by automatically applying the electric parking brake.
This week, the story is a little different, but still illustrates how software and hardware can interact unpredictably, especially in the automotive space. The story centers on a 2015 Optima recall for a software update for the knock sensor detection system. We can’t find the specifics, but if this recall on a similar Kia model in the same model year range and a class-action lawsuit are any indication, the update looks like it would have made the KSDS more sensitive to worn connecting rod damage, and forced the car into “limp home mode” to limit damage to the engine if knocking is detected.
A clever solution to a mechanical problem? Perhaps, but because the Kia owner in the story claims not to have received the snail-mail recall notice, she got no warning when her bearings started wearing out. Result: a $6,000 bill for a new engine, which she was forced to cover out of pocket. Granted, this software fix isn’t quite as egregious as Ford’s workaround for weak driveshaft mounting bolts, and there may very well have been a lack of maintenance by the car’s owner. But if you’re a Kia mechanical engineer, wouldn’t your first instinct have been to fix the problem causing the rod bearings to wear out, rather than papering over the problem with software?
Some people loved it, and a lot of space nerds really hated it. Either way, it’s hard to argue with the fact that NASA’s Space Shuttle program resulted in some pretty amazing engineering, and that artifacts from the program are highly sought-after as museum pieces. And perhaps no artifact is more highly prized than the three remaining orbiters, one of which will soon be at the center of a unique and awe-inspiring display. Endeavour (OV-105), which was moved to the California Science Center in Los Angeles in 2012 to much well-deserved fanfare, is going to be displayed in a full-stack, upright configuration. The orbiter will be reunited with external tank ET-94 and a pair of solid rocket boosters for the first time since she last flew in 2011. The 20-story stack will be erected over the next six months or so, after which a new wing of the museum will be built around it. It’ll be the first time a Space Shuttle full-stack assembly will be attempted outside of a NASA facility. We’re keen to see how this whole thing goes together — it’ll make for some interesting crane work.
Also in space news, we just learned that a US Senate subcommittee slashed the budget for the Mars Sample Return mission. And dramatically so; NASA wanted $949 million to fund fiscal year 2024 work on the mission, but senators offered just a third of that. NASA has already spent $1.74 billion on the Sample Return mission, which was supposed to fly in 2028 but looks like it’ll slip, drawing enough of the lawmakers’ ire for them to threaten to reallocate the $300 million to Artemis if NASA can’t prove the whole Sample Return mission won’t cost more than $5.3 billion. We’ve covered some of the ideas NASA is batting around for collecting and returning the sample tubes Perseverance is leaving on the surface of Mars; looks like they might have to scale things back a bit.
Is this it? Are we finally in the flying car future? Judging by the buzz recently, we’d be inclined to say yes, that is if we didn’t live in a reality-based world with multiple decades of observing just how poorly humans handle maneuvering a vehicle in only two dimensions. But still, we’re seeing a lot of stories about the FAA’s approval of the Alef Aeronautics Model A flying car. It’s important to note that this certification is for testing only, so it’s not like these things will be rolling flying off the assembly line anytime soon. Also in the zeitgeist is the in-no-way-sketchy ASKA A5 flying car, about which we can add little by way of comment to what you’ll see in the video below.
And finally, from the “If it ain’t broke, for the love of God, don’t touch it!” department, we were tipped off to a factory in Japan that’s relying on a 40-year-old computer to design their textiles. We couldn’t find any technical details, and the machine doesn’t seem to be any of the usual suspects from that era, although the keyboard layout and cassette drive bear a passing resemblance to a Commodore PET. The machine is shown only briefly and seems to run custom software for designing fabric patterns, which are stored on cassette tape and transferred to punch tape sheets for a Jacquard loom.
The rest of the video is a fascinating look inside an industry that is usually so optimized and scaled to ludicrous levels that watching the machines work is unsatisfying. Here you can actually see what’s going on with the looms that make the fabric for hanten, which are traditional winter coats. The handwork needed to manufacture these garments, which look remarkably cozy, is also a treat, especially the part where the padding is inserted. Enjoy the show.
I don’t think that computer is a PET. It used Sharp monitor and PET all used built in monitor with no logo on the CRT bezel
It’s a Sharp MZ-80K. Introduced in 1978, Z80 CPU, 48K of RAM (32K user-accessible).
I’m kind of amazed nobody else knew that, it was instantly recognisable just from the thumbnail
That one was never sold in USA as they couldn’t compete against PET so most American never knew of it.
It’s also written on the computer :)
around the 42 sec mark if anyone wants to look for it.
If you want a Fail for “We can fix it with software” you have to talk about the 737 Max. They took a certifiable aircraft and turned it into a death machine.
$100 million plane and a hundred lives ruined due to unreliable software.
Still did not have redundant sensors or properly trained pilots. There was an opportunity to shut the system off, had they known of it or how it operated.
It didn’t have triple redundant sensors but it did have dual redundant sensors. The sensors were the same as all the other 737s. Figuring out which one to trust is something a human can do if there is a failure or just ignore it all together. The software couldn’t however.
Yes there was a way to switch off the software but the airlines didn’t deam it necessary to train them on it.
It was supposed to have dual redundant sensors, which would show have shown a warning indicator when the sensors disagreed. Due to a software bug only one sensor was used and no warning was displayed, unless an optional upgrade had been purchased. Boeing discovered the bug before either crash, and decided fixing it could wait for a future software upgrade.
“They took a certifiable aircraft and turned it into a death machine.”
They took a proven series of aircraft (in themselves a study in product cycle extension), made it bigger, and upsold it based on increased capacity and – critically – the need for very little additional pilot training.
https://www.cnn.com/2019/03/22/politics/boeing-737-manual/index.html
If the stories are true about pilot training in certain countries, it wouldn’t be unreasonable to assume that those pilots got litttle to no useful training at all in the new systems beyond a perfunctory slide show tour of the features.
Since Boeing didn’t include the system in the manuals for pilots it’s safe to say the only pilots that had training in the new system was Boeing’s test pilots.
Mid-way through testing Boeing increased the values the system used so it was much more aggressive in “correcting” the aircraft’s behavior. Those updated values reportedly never made it into the FAA’s paperwork for the system, so it could be said the pilots weren’t the only ones with a limited understanding of what the plane was doing.
The entire thing is a case study in “swiss cheese”
Starting with redesigning an older airframe
Fitting newer, more powerful engines that altered the center of gravity and response to thrust (pitch up)
“Overcoming” that with dubious software fix, poorly documenting, not training pilots or even mentioning it really
Janky Angle of Attack sensor with no redundancy on most airplanes, and optional (expensive) upgrade to dual redundancy and in-cockpit annunciators for conflicting data.
Lack of regulatory oversight by federal agency, reliance on manufacturer to self-certify instead of third-party audit and certification
Lack of documenting and propagating critical information on how to resolve the errors the couple/few times the error happened on previous, non-accident flights.
And so on.
At least aviation has a pretty good track record of transparency about these types of issues and (eventual) studied, publicly published reports in nitty-gritty detail, along with recommendations on how to resolve issue(s).
Hospitals, and healthcare though….
so what is the punishment for Boeing
I actually had to think about that. It’s unclear to me if FAA/DOT investigations result in criminal or civil lawsuits but I’m inclined to say no. Unless you can demonstrate gross negligence (I’m 100% not a lawyer, would love some thoughts on the matter) then I think the answer is “nothing” unless family members bro g suit. but even that is looking at it wrong. As explained the “fault” is profoundly multi-factorial and not just Boeing – it’s also severe federal agencies and many other organizations including the airlines of several or many other nations that elected not to fit the optional hardware and software . I think.
Don’t you think Boeing has suffered enough with all the canceled orders?
The “optional”expensive upgrade to check for angle of attack disagreement was due to a software bug. The feature was designed and marketed to come standard on every plane (and quite rightfully so) but ended up being tied to an optional upgrade that would also display the sensor value.
Boeing knew about the bug for over a year before the first crash. After the first crash it was reviewed and the conclusion was … that it wasn’t an issue.
This one wasn’t on the airlines flying the 738 max, or regulatory authorities. Boeing had every chance to fix it, from different approaches to the max design early on right up until two planes flew into the ground and the whole fleet was were grounded.
https://www.cnn.com/2019/04/30/business/boeing-737-max-alert/index.html
https://www.npr.org/2019/05/06/720553748/boeing-knew-about-737-max-sensor-problem-before-plane-crash-in-indonesia
I can’t cite it so take it for what it’s worth.
“This one wasn’t on the airlines flying the 738 max, or regulatory authorities. ”
Not solely but airlines elected to not purchase The Thing, right or wrong. They elected not to train pilots on new aircraft. They made cost-benefit decisions that contributed to (not necissarily caused) the accidents.
Oversight of Boeing is on the US federal government and presumably other governments that certify the aircraft to operate within their borders and by their carriers. Taking Boeing’s word for it, right or wrong, is still on them somewhat.
My understanding is based on the official accident report and the TV show Air Disasters. Also take form that what you will.
It is multi factorial is what I’m saying and the definition of multiple small oversights leading to a disaster. Swiss Cheese.
Craig, I do agree there are multiple factors, and multiple opportunities missed to prevent the disaster by different parties. But it seems to me that the slice of cheese belonging to Boeing was mostly holes while the slices for governments and airlines are mostly cheese.
I’m basing that mostly on the reporting in the immediate aftermath of the accidents, though. Will have to take a look at the accident reports that are out now.
Looks like a Sharp MZ-80K
The computer in the factory video looks very similar to SHARP MZ-80K
As can be seen for example at: https://www.old-computers.com/history/detail.asp?n=50&t=3
and http://historyofpersonalcomputing.com/wordpress/sharp-mz-80k/
Best regards,
A/P Daniel F. Larrosa
Montevideo – Uruguay
They’re mad that the Sample Return Mission is behind schedule, so they’re going to give the money to Artemis??
Yeah, that makes sense.
The only reason the whole SLS/Orion/Artemis program exists is because powerful politicans were scared of the backlash from the job losses that would happen in shuttle factories when the shuttle program ended so they basically forced NASA to come up with something new that kept those factories in business and those workers in jobs. Its not surprising those same politicians want to keep wasting money on that junk.
Artemis = one and done.
Starship = failure after failure.
The exterior of ASKA sure is busy (i.e. ugly)!
The model and brand of the old computer are clearly visible in the low right corner of the video: as some others said, it is a Sharp, from the MZ-80K serie, but more precisely an MZ-80K2. And according to this site: https://monochromeeffect.org/JVCC/2020/08/22/sharp-mz-80k2/
“(…)the MZ-80K was sold as a DIY-kit for you to assemble. The MZ-80K2 was not sold as a kit, but rather a low-end version of the MZ-80B, and resembles the MZ-80K in a lot of ways, including the keyboard.”
And this other site: https://www.old-computers.com/museum/computer.asp?c=174
gives says it was released in 1980.